Bilge Pump Testing

Test a bilge pump by making it move real water against its real plumbing — never by listening for the motor. A pump that spins tells you nothing; a pump that lifts a rising bilge out through its own hose and skin fitting, with the float switch tripping on its own, tells you the safety system will actually clear water when the boat is taking it on. Bilge pumps are the one system aboard whose failures are completely silent until the emergency, so the only honest test is a wet one.

Why the maths says you are already behind

Start with the threat, because it sets the bar every pump has to clear. Water entering a hole below the waterline is governed by Torricelli's law: the efflux velocity is v = √(2gh), where h is the depth of the hole below the surface. Volumetric ingress is Q = C_d · A · √(2gh), with the discharge coefficient C_d around 0.6 to 0.65 for a sharp-edged orifice such as a failed skin fitting or a punched hole. The velocity term scales only with the square root of depth — so a hole a metre down is not twice as bad as one at 250 mm, only about twice — but it scales linearly with area, so diameter is everything.

The numbers are brutal. A 50 mm (roughly 2 inch) hole 300 mm below the waterline admits about 300 litres per minute — near 4,000 gph. A 100 mm hole at the same depth admits around 1,100 litres per minute, enough to founder a 30-foot yacht in minutes. A failed 1½-inch seacock three feet down floods at roughly 4,600 gph. Set that against a pump whose rated figure is 2000 gph and whose installed figure is well under half of it, and the arithmetic is unforgiving: no realistic bilge pump out-pumps a serious hull breach. The pump's honest job is to buy time against a modest leak — a weeping shaft seal, a cracked hose, a nuisance ingress around a keel ram or rudder gland — long enough to find and stem the source. That is the whole point of testing: you are not proving the pump saves the boat from anything, you are proving it delivers the real flow you are counting on, when you are counting on it.

What the flow-rate number really means

Rated gallons-per-hour is a bench figure taken at zero static head, open discharge, and roughly 13.6 volts — a charging-voltage test with nothing in the way. Afloat, none of those conditions holds, and the losses compound rather than add. Independent testing by the BoatUS Foundation (laboratory work at the University of Virginia) found submersible pumps delivering 15 to 33 percent below their rated capacity even close to ideal conditions — a 2000 gph unit measured near 1,500 gph at 12 volts before any real plumbing was in the way. Stack the installed losses on top:

• Static head (vertical lift). A centrifugal pump trades flow for head along its characteristic curve, and for typical aperture-impeller bilge pumps the working rule is about 8 percent of rated output lost per foot of lift. A Rule 2000 delivers on the order of 1,620 gph against one metre and about 1,300 gph against two metres of head — a 20 to 35 percent haircut before friction. Small units at six feet of lift produce under half their badge figure.
• Friction head (hose). Corrugated (convoluted) hose trips the flow into turbulence at every ridge and costs roughly 20 percent versus smooth-bore of the same nominal size. Horizontal run matters too: a useful approximation is that every 3 metres (about 10 feet) of horizontal hose adds around one foot of equivalent head. Elbows and tight radii add more.
• Restrictions (skin fitting). Nominal bore is not clear bore. The throat of a moulded 3/4-inch thru-hull can present roughly one-third the cross-sectional area of 3/4-inch hose, so the fitting alone behaves like several extra feet of head. Undersizing here silently throttles an otherwise healthy pump.
• Supply voltage. A settled, fully charged battery sits near 12.7 V — already about 6 percent below the 13.6 V test point — and sags further under the pump's inrush and the rest of the boat's load. Centrifugal output falls with voltage roughly as the square of speed, so a tired supply hits both flow and the maximum head the pump can even reach.

The practical upshot: a pump advertised at 2000 gph routinely clears well under half that once plumbed into a real boat. Size for the installed number, and — critically — measure it. This is the same "rated versus real" gap that catches people out with battery systems: the datasheet is a ceiling, not a working figure, and the working figure is the only one that clears water.

The wet-test method

Switching a pump on dry proves only that the motor turns. A dry centrifugal pump makes exactly the same sound whether it is clearing 1,500 litres an hour or nothing, and a dry run can hide an airlock, a partial blockage, a discharge disconnected out of sight, or a hose that has come adrift and is quietly recirculating water back into the bilge. Do it wet:

1. Flood the bilge with real water. Add enough clean water to submerge the intake and rise toward the float switch — you are recreating the event, not sprinkling the pump. See the note below on which pump types tolerate this and which must never be run dry.
2. Watch the discharge, not the panel. Station a crew member on deck and confirm a solid, steady stream leaves the skin fitting. A weak dribble against loud motor noise is the signature of head loss, blockage, an airlock, or air drawn in through a cracked hose or failing seal — the exact faults a dry test conceals.
3. Time the clear-down. Note how long the level takes to drop through a set distance. That is your installed flow baseline; a pump that clears the same bilge markedly slower than last regatta is degrading — a furring strainer, a stiffening impeller, rising circuit resistance — and it is telling you so in seconds.
4. Trip the float switch two ways. First lift it by hand to prove the switch and circuit; then let rising water lift it, proving the float moves freely, is not fouled on rope tails or grime, and triggers at a sensible level. Hand-testing alone passes a switch a real bilge would never reach.
5. Test the manual pump under a full stroke. Prime it and pump a genuinely flooded bilge to feel for a firm, consistent stroke and clean valve seating. A soft, spongy, or leaking action points to a perished diaphragm or a hardened valve flap — and confirm it delivers with the companionway and hatches shut, as offshore rules require.
6. Test every pump independently. Redundancy only counts if each pump clears water alone, on its own circuit and its own discharge. Switch one off and prove the other does the job by itself.

Log clear-down time, stream quality and switch behaviour so degradation reads as a trend, not a surprise. This is precisely the kind of item a pre-race inspection and a safety audit should capture in writing, not from memory.

Which pumps you may run dry, and which you must not

The old blanket rule "never run any bilge pump dry" is wrong, and knowing the difference protects the pump you are testing:

• Flexible (rubber) impeller pumps must never be run dry, even briefly. The vanes rely on the pumped water for lubrication and cooling; run dry they overheat, take a set, weld to the housing, and shear. Draw fuel or oil through one and the rubber swells and locks. Always prime these before testing.
• Diaphragm pumps are self-priming and tolerate running dry for extended periods, which is why they suit hard-to-flood or intermittently wet installations — but a strum box is still cheap insurance against a jammed valve.
• Submersible centrifugal pumps (the common aperture-impeller type) will tolerate short dry running, but the motor bearing and shaft seal are cooled and lubricated by the surrounding water; sustained dry cycling shortens seal and bearing life and is a leading cause of the seized-after-a-season failure. Test them submerged and do not leave them hunting dry against a chattering switch.

Float switches: types and how each one fails

The switch is the least reliable element of the automatic system, and it is worth knowing by mechanism which failure you are guarding against:

• Mechanical arm / ball switches pivot a float to close a micro-switch. Simple, but they jam on debris, fishing line and bilge grime, and the internal ball or contacts corrode. Failure modes: the pump never starts, or it runs continuously until the battery is flat — the latter being how a "reliable" boat is found with a dead bank and a burnt-out motor.
• Mercury-tilt switches are sealed and electrically robust but still depend on a free-moving float arm, so the mechanical fouling risk remains.
• Magnetic reed switches raise a magnet up a guide tube to close a sealed contact — no exposed electrics, but the float can foul the tube and stick.
• Electronic field-effect / capacitive probe switches sense water with no moving parts and shrug off mechanical jamming, but a film of oily bilge scum fools them: a conductive film can hold them on and drain the battery, a greasy insulating film can stop them switching off-to-on. A clean bilge is part of their maintenance, not optional.

Whatever the type, the switch must trip on rising water — not merely when poked — which is why the rising-water test above is non-negotiable. Where the rules and space allow, a high-water switch on a second pump wired to an audible alarm turns a silent failure into a warning.

Strainers, hoses and wiring — the failures around the pump

The pump is often the strongest link; the system around it fails first.

• Strum boxes and intakes collect debris and are the single most common cause of lost flow. Remove, clear and refit them at service, and keep the bilge clean so there is nothing to draw in. On a carbon boat, where sanding dust, cable ties and rope tails accumulate, this matters more, not less — the swarf that fairs a laminate is exactly the swarf that blinds a strainer.
• Discharge hose chafes where it crosses bulkheads and clamps, and corrugated hose splits at the ridges. Inspect the full run, double-clamp with all-316 stainless clamps at the pump and skin fitting, and — critically — route it in a high loop with an anti-siphon vent so a below-waterline outlet cannot back-siphon the sea into the boat through a stopped pump. This overlaps with the wider rope and hose wear discipline.
• Wiring and connections are the quiet killer. Salt-laden bilge air corrodes terminals and drives up resistance, dropping voltage at the motor and cutting flow. Use tinned, marine-grade cable with adhesive-lined heat-shrink crimp terminals, mounted high and dry; splices belong above any credible water level, not in the bilge. ABYC E-11 treats the bilge pump as a critical circuit, for which best practice holds voltage drop to 3 percent (about 0.36 V on a 12 V system, 0.72 V on 24 V) sized on the round-trip conductor length — get this wrong and the pump runs slow before anything is visibly faulty. Protect each pump with its own fuse, sited within about 7 inches (180 mm) of the supply, and the manual/auto/off switch on the same principle. A pump that "runs slow" is far more often a wiring, terminal or supply-voltage problem than a pump problem — measure the volts at the pump under load and look for the green bloom of salt corrosion before you condemn the motor.

What good looks like on a Grand Prix one-design

A stripped-out carbon Grand Prix boat sharpens every one of these points. On a Melges 40 there is essentially no accommodation — a bare carbon shell dominated by the structure and wet box around the electrically actuated Cariboni canting keel, with lines and rope tails run below deck. That keel is no small system: public class descriptions put a roughly 1.1-tonne bulb on a carbon fin of about 3.4 m, canting up to 45 degrees each side, driven by a single double-acting hydraulic ram off a dedicated high-power (order-of 24 V, several-kilowatt) Cariboni power pack with its own battery capacity. (Treat those figures as indicative and verify them against the class rules and the boat's documentation before relying on any.) The consequence for the bilge system is direct: a high-pressure hydraulic circuit and a moving keel penetration sit at the lowest point of a hull with almost nothing inside to slow water. A hydraulic hose weep, a ram-gland seep, or a keel-case ingress reports straight to the bilge, and the flat, open carbon pan lets it find the pump — and any exposed electrics — fast. Here the bilge system is genuine safety kit, not a convenience.

On a boat of this type, good practice looks like: pumps and switches mounted where they stay serviceable and cannot be fouled by stowed rope tails; the whole DC run kept high, tinned and corrosion-protected in a chronically damp carbon environment; a manual pump that is genuinely independent of the electrics and the keel batteries; discharge hose looped and anti-siphoned so a heeled, below-waterline outlet cannot feed the bilge; and a bilge kept scrupulously clean of the sanding dust and offcuts that plague carbon boats and blind strainers. Bad looks like a single automatic pump on a corroded circuit sharing the keel bank, a float buried under rope tails, no high-water alarm, and a "test" that consists of flicking the switch dockside.

Any boat-specific detail — the number, type and installed capacity of the pumps fitted, the discharge routing, and the required configuration for your safety category — must be verified against the class rules, the boat's own documentation, and the applicable World Sailing Offshore Special Regulations, not assumed. Treat this article as the method and the physics; take the numbers from the boat and the rules.

The takeaway

Bilge pumps are the system most likely to fail silently and among the most costly to have fail. The maths shows no realistic pump out-runs a serious breach, so its job is to buy time against a modest leak — which makes installed flow, not rated flow, the only figure that matters. Test them wet before every regatta and after any work that disturbs the system: confirm each pump lifts water through its own hose and skin fitting, cycle every float switch on rising water, clear the strum boxes, loop and anti-siphon the discharge, and check the wiring for the corrosion and voltage drop that starve a motor. Fold it into the pre-race inspection and safety audit, log clear-down times so degradation shows as a trend, and take the specific pump requirement from the Special Regulations and Notice of Race for the events you sail.

Bilge pump numbers, configuration and requirements must be verified against the boat's documentation, the class rules and the applicable World Sailing Offshore Special Regulations and the event's Notice of Race.